Know the blast radius

Every task is scored for what it could break.

Before a task runs, Clawvisor estimates its blast radius: how much damage it could do if it went wrong or was hijacked. Reading a calendar scores low and flows straight through. Deleting a production table scores high and waits for a human. Risk becomes a number you can gate on, not a gut feeling.
Blast radius scored

Drop and recreate the production analytics table.

data-agentexpires when task completes
Tools requested1

  • postgres.write

    DROP TABLE events

    write
high risk

The problem with treating every action the same

A flat approval flow can't tell the difference between an agent drafting a reply and an agent wiring money or dropping a database. So teams pick one extreme: review everything and grind to a halt, or review nothing and pray. The signal that should drive the decision, how irreversible and far-reaching the action is, never gets measured, so it never gets used.

How risk assessment works

01

Blast-radius scoring

Each task is scored on reversibility, scope, and sensitivity of the tools it requests, surfaced as a low / medium / high pill on the approval.

02

Risk-based gating

Set thresholds: auto-approve low risk, require a human above a line, hard-block the truly dangerous.

03

Context-aware

The same tool scores differently by scope: github.read on one repo is not github.write across the org.

04

Tuned to your tolerance

Adjust where the lines sit per team or environment, so staging moves fast and production stays careful.

Human attention lands on the few tasks that could actually hurt.

It's one of six power-ups you get from approving the task instead of the tools.

Fail-closed by default

Containment

Agents start with zero access. Every tool is granted per-task and revoked when the task ends. A compromised prompt can't reach what the task never needed.

See every action

Observability

A full audit trail records each tool call, argument, and decision, replayable and tied back to the task that authorized it.

Spend, per task

Cost attribution

Token and tool spend is attributed to the task and the human who approved it. Route cheaper models where the task allows it.

Agents without secrets

Credentials

Real secrets stay in the vault. Agents receive scoped, short-lived handles that Clawvisor swaps in at call time, never the credential itself.

Policy in plain English

Governance

Write RBAC/ABAC rules as readable policy. The same task can be auto-approved for one role and held for review for another.

Know the blast radiusYou're here

Risk assessment

Every task is scored for how much damage it could do if approved. Low-risk work flows straight through; high-blast-radius work waits for a human.

Let your agents fly. Safely.

Free to start for solo developers. Sign up and put a gateway in front of your agents in minutes. No credit card required.

Get startedRead the docs